M olecular MechanisM of vasculogenic MiMicry in tuMors
Mechanizm molekularny mimikry naczyniowej w nowotworach
Wen Li1, Jianrong Liu1, Jianlin Ren2, Qi Li3, Feng Gang Hou1
1Oncology Department of Shanghai Municipal Hospital of Traditional Chinese Medicine affiliated to Shanghai University of Traditional Chinese Medicine, Shanghai, China
2Scientific Research Office of Shanghai Municipal Hospital of Traditional Chinese Medicine affiliated to Shanghai University of Traditional Chinese Medicine, Shanghai, China
3Oncology Department of Shuguang Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai, China
Pielęgniarstwo Chirurgiczne i Angiologiczne 2015; 1: 22–27
Submitted: 26.02.2014, accepted: 19.01.2015 Corresponding author:
Feng Gang Hou PhD, Oncology Department of Shanghai Municipal Hospital of Traditional Chinese Medicine affiliated to Shanghai TCM University, 274 Zhijiang Road, Shanghai 200071, China, tel.: +86-21-56639828, fax: +86-21-56639484, e-mail: fghou555@126.com
Summary
Vasculogenic mimicry (VM) is a new model of tumor an- giogenesis, which refers to the de novo formation of perfus- able, matrix-rich, vasculogenic-like networks by virtue of the plasticity of aggressive tumor cells and the reconstruction of extracellular matrix (ECM). However, little is known about the molecular underpinnings of this phenomenon. To sum up, interactions of the critical VM-modulating pathways such as vascular (VE-cadherin, EphA2, VEGFR), embryonic and/or stem cell (Nodal, Notch4, CD133+, CD271), and hypoxia-relat- ed (HIF-1α, Twist1) signaling pathways are necessary for the formation of vasculogenic mimicry. What is more, accompa- nied by anti-angiogenic therapy of tumors, the use of VM is becoming more and more important.
Key words: vasculogenic mimicry (VM), molecular mech- anism, tumors.
Introduction
Tumor vasculogenic mimicry (VM), discovered in 1999, describes the de novo formation of perfusable, matrix-rich, vasculogenic-like networks by aggressive tu- mor cells in 3-dimensional matrices in vitro, which paral- lels matrix-rich networks in patients’ aggressive tumors.
The tumor cells capable of VM share a plastic phenotype and reconstruction of extracellular matrix (ECM). The molecular mechanism of VM is unclear. A large number of studies have contributed to understanding the molec- ular pathways supporting VM and indicate that endo- thelial pathways (VE-cadherin, EphA2, VEGFR), embryo or stem cell pathways (Nodal Notch4, CD133+, CD271), and hypoxia-related pathways (HIF-1 alpha, Twist1) play an important role in VM formation [1].
Formation of tumor VM
Phenotypic remodeling of the tumor cell
Phenotypic remodeling includes the reshaping of the genotype and the remodeling of morphological characteristics. In terms of genotype remodeling, Ma- niotis et al. found that 210 different genes are expres- sed in VM-positive and VM-negative human melanoma cells, including 15 genes associated with endothelial/
vascular phenotype [2]. Moreover, the molecular si- gnature of the tumor cell VM phenotype has revealed up-regulated expression of genes associated with em- bryonic progenitors, endothelial cells, vessel formation, matrix remodeling, and coagulation inhibitors, as well as down-regulation of genes predominantly associated with lineage-specific phenotype markers [3]. The epi-
Streszczenie
Mimikra naczyniowa (VM) to nowy model angiogenezy nowotworowej, który dotyczy tworzenia się de novo perfuzyj- nych sieci bogatych w macierz, zbliżonych do waskulogennych ze względu na plastyczność agresywnych komórek nowotwo- rowych oraz rekonstrukcję macierzy pozakomórkowej (ECM).
Wiedza na temat molekularnych podstaw tego zjawiska jest jednak niewielka. Reasumując – interakcje decydujących ka- skad modulujących mimikrę naczyniową, takich jak kaska- da naczyniowa (VE-kadheryna, EphA2, VEGFR), embrionalna i/lub komórek macierzystych (Nodal, Notch4, CD133+, CD271) oraz związana z hipoksją (HIF-1α, Twist1), są niezbędne do tworzenia się mimikry naczyniowej. Ponadto wraz z antyan- giogenną terapią nowotworów zastosowanie mimikry naczy- niowej staje się coraz ważniejsze.
Słowa kluczowe: mimikra naczyniowa, mechanizm mole- kularny, nowotwory.
thelial-mesenchymal transition (EMT) is a remodeling process whereby epithelial cells reduce epithelial cha- racteristics, such as a decrease of cell-cell contact and down-regulation of E-cadherin, while simultaneously acquiring mesenchymal properties including fibroblast- like shape, increased cell motility and up-regulation of mesenchymal markers such as vimentin and cadherin 5 (CDH5) [4]. Moreover, aberrant expression of EMT regu- lators was found in VM forming cancer cell lines, Twist1 in human hepatocellular carcinoma (HCC) cells [5], and ZEB1 in colorectal carcinoma cells [6]. In addition, hy- poxia can enhance VM capacity of HCC cells through increased Twist1 expression [7]. Collectively, these ac- cumulating findings provide supportive evidence that tumor cells capable of VM exhibit a high degree of phe- notypic plasticity.
Extracellular matrix reconstitution
The extracellular matrix reconstruction of tumors mainly takes place within cells, and tumor cells inte- ract with the matrix. The change of extracellular ma- trix components is advantageous to the formation of VM [8]. In the process of extracellular matrix remo- deling, abnormal composition of extracellular matrix degradation is the key step. The matrix metalloprote- inases (MMPs), tissue inhibitors of metalloproteinases (TIMPs) and original urokinase type fibrinolytic enzyme activator (uPA) are of great importance. MMPs increase tumor cell surface adhesion protein receptor and lami- nin adhesion, and make the LN-5γ2 chain split into γ2’
and γ2x to destroy the integrity of the extracellular ma- trix. At the same time, TIMPs block proteolytic activity and promote the matrix changing to a structure that is helpful to the formation of VM [9-11]. uPA is a serine proteolytic enzyme involved in activation of fibrinolytic enzyme. First of all, uPA combines with the urokinase fibrinolytic enzyme activator receptor (uPAR) to promo- te PLG into fibrinolytic enzyme (PLM), which degrades ECM, basement membrane protein fiber and protein.
Secondly, uPAR is not evenly distributed in tumor cell membranes, but is mainly in front of the contact area between ECM and basement membrane. As a result, the uneven distribution of uPAR and high concentra- tion of fibrinolytic enzyme help to orientate degrada- tion of ECM, making it easier for tumor cells to break the bondage of the matrix [12, 13]. Therefore, uPA plays an important role in tumor invasion, metastasis, angio- genesis and VM [14]. In addition, the apoptosis-related protein caspase-3 performs a non-apoptotic function to promote VM formation of melanoma cells. Inhibition of caspase-3 activity by using low-dose z-DEVD-fmk and down regulation of caspase-3 by specific small inter- fering RNA reduced VM formation of melanoma cells in vitro [15]. Caspase-3-mediated promotion of VM formation may be attributed to the cleavage of matrix metalloproteinase-2.
Molecular mechanisms of VM formation
The plasticity of aggressive tumor cells and the re- construction of ECM lead to the formation of VM. The molecular basis of this process is supported by the complex cooperation of signaling pathways; in parti- cular, critical VM modulating genes can be categorized into pathways associated with vascular, embryonic and/or stem cell, and hypoxia signaling. Although a my- riad of genes having a relationship with VM have been reported, interactions of these three pathways play the role of a platform (Figure 1).
Vascular signaling pathways
The first vascular-associated gene shown to be in- volved in VM is the VE-cadherin gene (CDH5). In 2006 Hess et al. showed that in VM melanoma VE-cadherin colocalizes with EphA2 at areas of cell-cell contact and that these two molecules are able to directly or indirec- tly interact during the process of VM [16]. VE-cadherin regulated erythropoietin-producing hepatocellular car- cinoma-A2 (EphA2) by mediating it to become phospho- rylated through interactions with its membrane-bound ligand ephrin-A1 [17]. Phosphorylated EphA2 subse- quently activates phosphoinositide 3-kinase (PI3K), upregulates matrix metalloproteinase (MMP) 14 expres- sion, and activates MMP2. Both MMP14 and MMP2 promote cleavage of the laminin-5γ2 chain into γ2’ and γ2x, which in turn stimulate migration, invasion, and VM in melanoma cells [18]. In addition, VE-cadherin expression and activity are enhanced by binding of the transcription factor Twist1 to the VE-cadherin promoter, whereas down-regulation of Twist1 expression leads to decreased VE-cadherin, MMP2, and MMP9 expression and VM formation in human hepatocellular carcinoma cells [5]. Recently, Bcl-2 has also been shown to indu- ce VM in human melanoma cells. Bcl-2 over-expression increased VE-cadherin expression and VM formation under normoxia, whereas Bcl-2 siRNA significantly re- duced VE-cadherin expression and VM formation under hypoxia. Therefore, Bcl-2-dependent VE-cadherin over- -expression may be an important mechanism by which hypoxia induces VM [19].
In recent years, vascular endothelial growth factor- -A (VEGF-A), a well-characterized promoter of endothe- lial cell proliferation, survival, and angiogenesis, has been linked to VM in both melanoma and ovarian car- cinoma. In melanoma, the autocrine secretion of VEGF- -A is required for VM, largely through activation of VEGF receptor 1 (VEGFR1). VM appears to be mediated thro- ugh the activation of PI3K/PKC downstream of VEGFR1, in co-operation with integrin-mediated signaling path- ways in melanoma [20-22]. Addition of VEGF-A to ova- rian carcinoma cells promoted the up-regulation of VM- -associated genes, including the genes for VE-cadherin,
EPHA2, MMP-2, and MMP-9. Furthermore, it has also been demonstrated that the interaction of EPHA2 and VEGF promotes tumor cell plasticity essential for VM [23]. COX-2, an enzyme responsible for catalyzing the conversion of arachidonic acid into primarily pro- staglandin E2 (PGE2), has been found to increase the expression of PGE2, and has also been found to incre- ase the expression of VEGF through a protein kinase C (PKC) mediated pathway. PGE2 binds to a family of pro- stanoid receptors (prostaglandin E2 receptor subtypes EP1-4), which in turn activate EGF receptor (EGFR) si- gnaling, and PKC-dependent ERK1/2 activation signa- ling [24]. COX-2/PGE2/EP3 promotion of VM via VEGF has been shown in human pancreatic cancer cells [25].
Recently, it has also been demonstrated that COX-2/
EP4 controls tumor growth, angiogenesis, lymphangio- genesis and metastasis of the lungs and lymph nodes in a breast cancer model [26].
VEGF receptor tyrosine kinases (VEGFR1 and VEGFR2) bind VEGF-A in an autocrine or paracrine manner and the activation of the PI3K/Akt pathway, whereas cancer cell invasion and migration have been shown to involve VEGFR1 activation of Src and ERK1/2 pathways [27]. VEGFR1, but not VEGFR2, mediates VEGF-A-induced VM in melanoma cells, and it has been postulated that VM is mediated through the synergistic transduction of VEGF-A/VEGF R1/ PI3K/ PKC and inte-
grin-signaling pathways [22]. However, VEGFR-2 (Flk-1) plays a key role in VM formation of glioblastoma-deri- ved tumor cells via activation of focal adhesion kinase (FAK) and mitogen-activated protein kinase ERK1/2. In contrast, blockade of VEGF activity by the neutralizing antibody bevacizumab fails to recapitulate the impact of VM formation, suggesting that Flk-1-mediated VM is independent of VEGF [28]. However, glioma stem-like cells (GSLCs) preferentially expressed VEGFR-2 that must depend on activating VEGF and then mediates VM [29]. In melanoma cell lines, endothelin-1 (ET-1) in combination with VEGF-C further increased VEGFR-3, MAPK, and AKT phosphorylation and markedly pro- moted cell migration and vasculogenic mimicry [30].
Although VEGF-C was identified as a lymphangiogenic growth factor and later shown to promote tumor meta- stasis, the relation with VM may be found soon.
Tissue factor (TF) pathway inhibitor 1 (TFPI-1) and 2 (TFPI-2) are critical genes that initiate and regulate the coagulation pathways. The procoagulant function of TF in highly aggressive melanoma is shown to be re- gulated by TFPI-1 but not by TFPI-2. Antibody inhibition experiments reveal that TFPI-2 is required for VM in vi- tro, but plasmin is an unlikely target protease of TFPI-2.
Blockade of TFPI-2 suppressed MMP-2 activation, and, therefore, TFPI-2 appears to regulate an essential pa- thway of VM. Culturing poorly aggressive melanoma Fig. 1. Pathways associated with vascular, embryonic and/or stem cell, and hypoxia signaling
cells on three-dimensional matrix containing recombi- nant TFPI-2 produces some of the phenotypic change associated with aggressive, vasculogenic melanoma cells. Thus, TFPI-2 contributes to VM plasticity, where- as TFPI-1 has anticoagulant functions of relevance for perfusion of VM channels formed by TF-expressing me- lanoma cells [31].
Galectin-3 (Gal-3) is a beta-galactoside-binding protein that is involved in cancer progression and me- tastasis. In vitro, Gal-3 silencing results in loss of tu- mor cell invasiveness and capacity to form VM. cDNA microarray analysis after Gal-3 silencing revealed that Gal-3 regulates the expression of multiple genes, inc- luding VE-cadherin, interleukin-8 (IL-8), fibronectin-1, endothelial differentiation sphingolipid G-protein re- ceptor-1, and MMP-2. Chromatin immunoprecipitation assays and promoter analyses revealed that Gal-3 si- lencing resulted in a decrease of VE-cadherin and IL-8 promoter activities due to enhanced recruitment of transcription factor early growth response-1 (EGR-1).
Moreover, transient over-expression of early growth re- sponse-1 in C8161 cells resulted in a loss of VE-cadhe- rin and IL-8 promoter activities and protein expression [32, 33]. Thus, Gal-3 plays an essential role during the acquisition of vasculogenic mimicry and angiogenic properties associated with melanoma progression.
Embryonic and/or stem cell pathways
Nodal signaling pathways are important regula- tors of human embryonic pluripotency and vertebrate embryonic development [34]. Nodal is a growth factor of the TGFβ superfamily that binds Cripto-1 and ac- tivates type I and type II activin-like kinase receptors (ALK4/5/7 and ActRIIB, respectively), which subsequ- ently propagates canonical signaling via Smad2/3 [35, 36]. Activated Smad2/3 regulate the formation of VM by the expression of VE-cadherin. Nodal expression is influenced by Notch, which consists of 4 transmem- brane receptors (Notch1-4), whose signaling pathways are critical regulators of vertebrate embryogenesis [36, 37]. Notch signaling is initiated by binding of a Notch ligand, which induces a series of cleavages that genera- te the release of the Notch intracellular domain (NICD).
The NICD translocates to the nucleus and regulates the expression of a number of context-dependent targets, including Nodal [38, 39]. Notch4 functions primarily in vascular development and is enriched in the subpopu- lation of melanoma cells that form VM [40]. Moreover, in glioblastoma stem-like cells, melanoma cells, triple- -negative (TN) breast cancer cells and bladder urothe- lial carcinoma cells, the stem cell marker CD133+ plays a pivotal role in the formation of VM [41-43]. In addi- tion, one of the most studied tumor stem cell markers, CD271 (also known as nerve growth factor receptor, NGFR or p75NTR), is a neurotrophin receptor, which can
bind all of the neurotrophins with similar affinity [44].
It was found that the VM-forming uveal melanoma cell lines in 3D cultures expressed CD271. In contrast, cells grown in 2D cultures and tumor cell subpopulations not participating in VM formation in 3D cultures were negative for CD271 [45]. All the data demonstrate that the embryonic and/or stem cell pathway is very impor- tant in VM.
Hypoxia-related signaling pathways
Hypoxia is a hallmark of most solid tumors and can regulate pathways in VM formation. The hypoxia-in- ducible factor (HIF) complex (composed of HIF-1 and one HIF-subunit: HIF-1, HIF-2, or HIF-3) is a key regu- lator of oxygen homeostasis in both physiological and pathological environments. Recently, hypoxia has also been shown to induce VM in hepatocellular carcinoma, Ewing’s sarcoma, and melanoma [46]. Pertinent to VM, hypoxia can directly modulate VEGF-A, VEGFR1, EPHA2, Twist1, Nodal, osteopontin, and COX-2 gene expression (via HIF/hypoxia response element binding) or indirec- tly modulate VE-cadherin, TF, and PEDF expression (via activation of an intermediary protein that regulates gene transcription or post-transcriptional protein pro- cessing) [47, 48]. In addition, hypoxia can modulate the formation of VM via HIF-1α in human gallbladder carcinoma [49]. Moreover, in the human fibrosarcoma- -derived cell line HT1080, the angiogenic marker neu- ropilin-1 (NRP-1) also plays a vital role in VM formation.
Experiments performed with HT1080 cells stably trans- fected with plasmid constructs expressing shNRP-1 or full-length NRP-1 clearly established that hypoxia- -mediated HIF-1α-dependent up-regulation of NRP-1 is a critical molecular event involved in VM [50]. Reactive oxygen species (ROS) regulate VM formation by stabili- zing the regulation of HIF-1α and increasing the sensi- tivity to hypoxia of melanoma cells [51]. Recent studies have shown that pVHL, the protein product of the VHL gene, adjusts the level of HIF-1α to control the expres- sion of VEGF, VE-cadherin, EphA2 and regulation of VM [52, 53].
Anti-tumor research on VM
Although the use of antiangiogenic compounds to target the blood supply of a tumor seems logical, the success of these compounds in the clinic has been very limited. Bevacizumab, sorafenib, and sunitinib target the transdifferentiated endothelial phenotype (VEGF, VEGFRs, PDGFR, RET) with limited efficacy. Clinical use of these drugs has been shown to limit the growth of the primary tumor, but long-lasting effects are rare and typically lead to only moderate benefits for overall su- rvival [54, 55]. There is experimental evidence that spe-
cifically targeting pathways implicated in VM may have success in inhibiting tumor growth. Certainly, a handful of preclinical studies suggest that specific compounds affecting components of the previously described va- scular, embryonic, or hypoxia on VM. However, there are only two drugs, the FDA-approved angiogenesis in- hibitors thalidomide and rapamycin, that are proved to inhibit VM formation [23]. There are many medicines inhibiting vascular mimicry, which are still in the expe- rimental stage. When in the murine choroidal melano- ma model animals were given curcumin once a day at a dose of 100 mg/kg for 18 days, the results indicated that the tumor volume was reduced and the numbers of VM, mosaic vessels, and endothelium-dependent vessels were significantly decreased compared with the control group [56]. That is to say, curcumin has the ability to inhibit the growth of engrafted melanoma VM channels through the regulation of vasculogenic factors that could be related to the down-regulation of the EphA2/PI3K/MMPs signaling pathway. Thus, curcu- min has the potential of being a clinical inhibitor of VM of choroidal melanoma [57]. Moreover, the inhibitory effect of nicotinamide on VM formation could be at le- ast partially explained by nicotinamide-driven downre- gulation of VE-cadherin [58]. More studies have found that not only can lycorine hydrochloride (LH) inhibit the metastatic melanoma cell line C8161 VM structure, but also the VM structure and the volume of transplanta- tion tumor in nude mouse can be inhibited [59]. Howe- ver, these studies are still at the experimental stage, and the most effective anti-cancer drugs for VM still need more in-depth research.
References
1. Maniotis AJ, Folberg R, Hess A, et al. Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry.
Am J Pathol 1999; 155: 739-752.
2. Hendrix MJ, Seftor EA, Hess AR, Seftor RE. Vasculogenic mimicry and tumour-cell plasticity: lessons from melanoma. Nature Re- views Cancer 2003; 3: 411-421.
3. Hess AR, Seftor EA, Gardner LM, et al. Molecular regulation of tumor cell vasculogenic mimicry by tyrosine phosphorylation role of epi- thelial cell kinase (Eck/EphA2). Cancer Res 2001; 61: 3250-3255.
4. Lirdprapamongkol K, Chiablaem K, Sila-Asna M, et al. Exploring stemness gene expression and vasculogenic mimicry capacity in well-and poorly-differentiated hepatocellular carcinoma cell lines.
Biochem Biophys Res Com 2012; 422: 429-435.
5. Sun T, Zhao N, Zhao XL, et al. Expression and functional significan- ce of Twist1 in hepatocellular carcinoma: its role in vasculogenic mimicry. Hepatology 2010; 51: 545-556.
6. Liu Z, Sun B, Qi L et al. Zinc finger E-box binding homeobox 1 promotes vasculogenic mimicry in colorectal cancer through in- duction of epithelial-to-mesenchymal transition. Cancer Science 2012; 103: 813-820.
7. Ma JL, Han SX, Zhu Q, et al. Role of Twist in vasculogenic mimicry formation in hypoxic hepatocellular carcinoma cells in vitro. Bio- chem Biophys Res Commun 2011; 408: 686-691.
8. Huttenlocher A. ECM: chemoattraction but not adhesion. Blood 2013; 121: 1489-1491.
9. Lai K, Conway R, Crouch R, et al. Expression and distribution of MMPs and TIMPs in human uveal melanoma. Exp Eye Res 2008;
86: 936-941.
10. Handsley MM, Edwards DR. Metalloproteinases and their inhibi- tors in tumor angiogenesis. Int J Can 2005; 115: 849-860.
11. Bu Z, Pan Y, Shang B, et al. SZ-117, a monoclonal antibody against matrix metalloproteinase-2 inhibits tumor cell-mediated angioge- nesis. Hybridoma 2012; 31: 63-67.
12. Andreasen PA, Kjøller L, Christensen L, Duffy MJ. The urokinase-ty- pe plasminogen activator system in cancer metastasis: a review.
Int J Can 1997; 72: 1-22.
13. Gokhale A, Kunder R, Goel A, et al. Distinctive microRNA signature of medulloblastomas associated with the WNT signaling pathway.
Journal of cancer research and therapeutics 2010; 6: 521.
14. Parri M, Taddei ML, Bianchini F, et al. EphA2 reexpression prompts invasion of melanoma cells shifting from mesenchymal to amo- eboid-like motility style. Cancer Res 2009; 69: 2072-2081.
15. Liu YR, Sun B, Zhao XL, et al. Basal caspase-3 activity promotes migration, invasion, and vasculogenic mimicry formation of mela- noma cells. Melanoma Res 2013; 23: 243-253.
16. Hess AR, Seftor EA, Gruman LM, et al. VE-cadherin regulates EphA2 in aggressive melanoma cells through a novel signaling pathway: implications for vasculogenic mimicry. Cancer Biol Ther 2006; 5: 228-233.
17. Hess AR, Margaryan NV, Seftor EA, Hendrix MJ. Deciphering the signaling events that promote melanoma tumor cell vasculogenic mimicry and their link to embryonic vasculogenesis: role of the Eph receptors. Dev Dyn 2007; 236: 3283-3296.
18. Lu XS, Sun W, Ge CY, et al. Contribution of the PI3K/MMPs/Ln-5γ2 and EphA2/FAK/Paxillin signaling pathways to tumor growth and vasculogenic mimicry of gallbladder carcinomas. Int J Oncol 2013;
42: 2103-2115.
19. Zhao N, Sun BC, Sun T, et al. Hypoxia-induced vasculogenic mi- micry formation via VE-cadherin regulation by Bcl-2. Med Oncol 2012; 29: 3599-3607.
20. Vartanian A, Stepanova E, Grigorieva I, et al. VEGFR1 and PKCα signaling control melanoma vasculogenic mimicry in a VEGFR2 kinase-independent manner. Melanoma Res 2011; 21: 91-98.
21. Wang JY, Sun T, Zhao XL, et al. Functional significance of VEGF-a in human ovarian carcinoma: role in vasculogenic mimicry. Cancer Biol Ther 2008; 7: 758-766.
22. Frank NY, Schatton T, Kim S, et al. VEGFR-1 expressed by mali- gnant melanoma-initiating cells is required for tumor growth.
Cancer Res 2011; 71: 1474-1485.
23. Seftor RE, Hess AR, Seftor EA, et al. Tumor cell vasculogenic mimi- cry: from controversy to therapeutic promise. Am J Pathol 2012;
181: 1115-1125.
24. Wu WK, Yiu Sung JJ, Lee CW, et al. Cyclooxygenase-2 in tumori- genesis of gastrointestinal cancers: an update on the molecular mechanisms. Cancer Lett 2010; 295: 7-16.
25. Ma JX, Sun YL, Wang YQ, et al. Triptolide induces apoptosis and inhibits the growth and angiogenesis of human pancreatic can- cer cells by downregulating COX-2 and VEGF. Oncol Res 2013; 20:
359-368.
26. Xin X, Majumder M, Girish GV, et al. Targeting COX-2 and EP4 to control tumor growth, angiogenesis, lymphangiogenesis and me- tastasis to the lungs and lymph nodes in a breast cancer model.
Lab Invest 2012; 92: 1115-1128.
27. Claesson-Welsh L. Signal transduction by vascular endothelial growth factor receptors. Biochem Soc Trans 2003; 31: 20-24.
28. Francescone R, Scully S, Bentley B, et al. Glioblastoma-derived tu- mor cells induce vasculogenic mimicry through Flk-1 protein acti- vation. J Biol Chem 2012; 287: 24821-24831.
29. Yao X, Ping Y, Liu Y, et al. Vascular endothelial growth factor recep- tor 2 (VEGFR-2) plays a key role in vasculogenic mimicry forma- tion, neovascularization and tumor initiation by glioma stem-like cells. PloS One 2013; 8: e57188.
30. Spinella F, Caprara V, Di Castro V, et al. Endothelin-1 induces the transactivation of vascular endothelial growth factor receptor-3 and modulates cell migration and vasculogenic mimicry in mela- noma cells. J Mol Med 2013; 91: 395-405.
31. Ruf W, Seftor EA, Petrovan RJ, et al. Differential role of tissue factor pathway inhibitors 1 and 2 in melanoma vasculogenic mimicry.
Cancer Res 2003; 63: 5381-5389.
32. Mourad-Zeidan AA, Melnikova VO, Wang H, et al. Expression pro- filing of Galectin-3-depleted melanoma cells reveals its major role in melanoma cell plasticity and vasculogenic mimicry. Am J Pathol 2008; 173: 1839-1852.
33. Kim ES, Lim DJ, Lee K, et al. Absence of galectin-3 immunostaining in fine-needle aspiration cytology specimens from papillary thy- roid carcinoma is associated with favorable pathological indices.
Thyroid 2012; 22: 1244-1250.
34. Schier AF, Shen MM. Nodal signalling in vertebrate development.
Nature 2000; 403: 385-389.
35. Postovit LM, Seftor EA, Seftor RE, Hendrix MJ. Targeting Nodal in malignant melanoma cells. Expert Opin Ther Targets 2007; 11:
497-505.
36. Wakefield LM, Hill CS. Beyond TGFβ: roles of other TGFβ superfa- mily members in cancer. Natur Rev Cancer 2013; 13: 328-341.
37. Strizzi L, Hardy KM, Seftor EA, et al. Lessons from embryogenesis.
In Melanoma Development. Springer 2011; 281-296.
38. McAllister JC, Zhan Q, Weishaupt C, et al. The embryonic morpho- gen, Nodal, is associated with channel-like structures in human malignant melanoma xenografts. J Cutan Pathol 2010; 37: 19-25.
39. Vartanian A, Gatsina G, Grigorieva I, et al. The involvement of Notch signaling in melanoma vasculogenic mimicry. Clin Exp Med 2013; 13: 201-209.
40. Strizzi L, Hardy KM, Margaryan NV, et al. Potential for the embry- onic morphogen Nodal as a prognostic and predictive biomarker in breast cancer. Breast Can Res 2012; 14: R75.
41. Chiao MT, Yang YC, Cheng WY, et al. CD133+ glioblastoma stem- like cells induce vascular mimicry in vivo. Curr Neurovasc Res 2011; 8: 210-219.
42. Liu T, Sun B, Zhao X, et al. CD133+ cells with cancer stem cell characteristics associates with vasculogenic mimicry in triple-ne- gative breast cancer. Oncogene 2013; 32: 544-553.
43. Yu L, Wu S, Zhou L, et al. [Expressions of CD133 and CD82/KAI1 in bladder urothelial carcinoma and their correlation with vascu- logenic mimicry]. Nan Fang Yi Ke Da Xue Xue Bao 2013; 33: 1336- 1340.
44. Micera A, Lambiase A, Stampachiacchiere B, et al. Nerve growth factor and tissue repair remodeling: trkA(NGFR) and p75(NTR), two receptors one fate. Cytokine Growth Factor Rev 2007; 18:
245-256.
45. Valyi-Nagy K, Kormos B, Ali M, et al. Stem cell marker CD271 is expressed by vasculogenic mimicry-forming uveal melanoma cells in three-dimensional cultures. Mol Vis 2012; 18: 588-592.
46. Kirschmann DA, Seftor EA, Hardy KM, et al. Molecular pathways:
vasculogenic mimicry in tumor cells: diagnostic and therapeutic implications. Clin Cancer Res 2012; 18: 2726-2732.
47. Fernández-Barral A, Orgaz JL, Gomez V, et al. Hypoxia negative- ly regulates antimetastatic PEDF in melanoma cells by a hypoxia inducible factor-independent, autophagy dependent mechanism.
PloS One 2012; 7: e32989.
48. Mao XG, Xue XY, Wang L, et al. CDH5 is specifically activated in glioblastoma stemlike cells and contributes to vasculogenic mimi- cry induced by hypoxia. Neuro Oncol 2013; 15: 865-879.
49. Sun W, Shen ZY, Zhang H, et al. Overexpression of HIF-1α in prima- ry gallbladder carcinoma and its relation to vasculogenic mimicry and unfavourable prognosis. Oncology reports 2012; 27: 1990- 2002.
50. Misra RM, Bajaj MS, Kale VP. Vasculogenic Mimicry of HT1080 Tu- mour Cells In Vivo: Critical Role of HIF-1α-Neuropilin-1 Axis. PloS One 2012; 7: e50153.
51. Comito G, Calvani M, Giannoni E et al. HIF-1α stabilization by mi- tochondrial ROS promotes Met-dependent invasive growth and vasculogenic mimicry in melanoma cells. Free Radic Biol Med 2011; 51: 893-904.
52. Berndt JD, Moon RT, Major MB. β-catenin gets jaded and von Hip- pel-Lindau is to blame. Trends Biochem Sci 2009; 34: 101-104.
53. Franco CA, Liebner S, Gerhardt H. Vascular morphogenesis: a Wnt for every vessel? Curr Opin Genet Dev 2009; 19: 476-483.
54. Bergers G, Hanahan D. Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer 2008; 8: 592-603.
55. Conley SJ, Gheordunescu E, Kakarala P, et al. Antiangiogenic agents increase breast cancer stem cells via the generation of tumor hypo- xia. Proc Nat Acad Sci USA 2012; 109: 2784-2789.
56. Noguera-Troise I, Daly C, Papadopoulos NJ, et al. Blockade of Dll4 inhibits tumour growth by promoting non-productive angiogene- sis. Nature 2006; 444: 1032-1037.
57. Chen LX, He YJ, Zhao SZ, et al. Inhibition of tumor growth and vasculogenic mimicry by curcumin through down-regulation of the EphA2/PI3K/MMP pathway in a murine choroidal melanoma model. Cancer Biol Ther 2011; 11: 229-235.
58. Itzhaki O, Greenberg E, Shalmon B, et al. Nicotinamide inhibits vasculogenic mimicry, an alternative vascularization pathway ob- served in highly aggressive melanoma. PloS One 2013; 8: e57160.
59. Liu R, Cao Z, Tu J, et al. Lycorine hydrochloride inhibits metastatic melanoma cell-dominant vasculogenic mimicry. Pigment Cell Me- lanoma Res 2012; 25: 630-638.
